Intelligent conversational agent as an active strategy for school safety management

Unsafe educational environments jeopardize the well-being of children and adolescents and negatively impact their educational results. This study presents the development and evaluation of an AI-based platform designed to enhance school safety management and support the systematic creation and dissemination of Comprehensive School Safety Plans (PISE) within Chilean educational institutions located in high-risk coastal areas. Using a participatory design thinking approach, the research involved school communities from two secondary schools in southern Chile. The platform comprises two main components: a Web application that facilitates collaborative development and a generative AI-powered conversational chatbot integrated into WhatsApp, which serves as an accessible communication channel to disseminate emergency plans and respond to community queries. The content analysis revealed significant gaps in existing emergency plans, particularly in terms of identifying territorial risks, protocols for specific emergencies, and communication strategies. The intelligent platform addressed these issues by streamlining the creation of plans, providing contextualized guidance, and enhancing user engagement. Interaction with the conversational chatbot revealed a wider range of perceived safety threats than those officially documented, including internal and external violence, health emergencies, and accidents, highlighting discrepancies between institutional records and community concerns. User feedback demonstrated high acceptance of the platform’s usability and utility, with the chatbot fostering anonymous, real-time dialogue that improved awareness and identified training needs. However, challenges such as response accuracy and system limitations were noted and addressed through iterative refinements. The study underscores the potential of AI-driven tools to improve traditional safety management by promoting participatory risk identification, improving communication, and supporting dynamic emergency planning tailored to local contexts. Further validation in diverse educational settings is recommended to generalize findings and optimize implementation strategies. This research contributes a novel model for integrating AI conversational agents in school safety, emphasizing ethical design and community involvement to safeguard student well-being and educational continuity.